Definitions of functions regarding COCO observers. More...

#include "coco.h"
#include "coco_internal.h"
#include <limits.h>
#include <float.h>
#include <math.h>
#include "logger_bbob.c"
#include "logger_biobj.c"
#include "logger_toy.c"
#include "logger_rw.c"

Data Structures
struct	coco_observer_targets_t
	The type for triggers based on target values. More...

struct	coco_observer_evaluations_t
	The type for triggers based on numbers of evaluations. More...

Macros
#define	COCO_MAX_EVALS_TO_LOG 1000
	The maximum number of evaluations to trigger logging. More...

Functions
static coco_observer_t *	coco_observer_allocate (const char result_folder, const char observer_name, const char algorithm_name, const char algorithm_info, const size_t number_target_triggers, const double target_precision, const size_t number_evaluation_triggers, const char *base_evaluation_triggers, const int precision_x, const int precision_f, const int precision_g, const int log_discrete_as_int)
	Allocates memory for a coco_observer_t instance. More...

void	coco_observer_free (coco_observer_t *observer)
	Frees the given observer. More...

coco_observer_t *	coco_observer (const char observer_name, const char observer_options)
	Constructs a COCO observer. More...

coco_problem_t *	coco_problem_add_observer (coco_problem_t problem, coco_observer_t observer)
	Adds an observer to the given problem. More...

coco_problem_t *	coco_problem_remove_observer (coco_problem_t problem, coco_observer_t observer)
	Removes an observer from the given problem. More...

const char *	coco_observer_get_result_folder (const coco_observer_t *observer)
	Returns result folder name, where logger output is written. More...

Methods regarding triggers based on target values
static coco_observer_targets_t *	coco_observer_targets (const size_t number_of_targets, const double precision)
	Creates and returns a structure containing information on targets. More...

static int	coco_observer_targets_trigger (coco_observer_targets_t *targets, const double given_value)
	Computes and returns whether the given value should trigger logging. More...

Methods regarding triggers based on numbers of evaluations.
static coco_observer_evaluations_t *	coco_observer_evaluations (const char *base_evaluations, const size_t dimension)
	Creates and returns a structure containing information on triggers based on evaluation numbers. More...

static int	coco_observer_evaluations_trigger_first (coco_observer_evaluations_t *evaluations, const size_t evaluation_number)
	Computes and returns whether the given evaluation number triggers the first condition of the logging based on the number of evaluations. More...

static int	coco_observer_evaluations_trigger_second (coco_observer_evaluations_t *evaluations, const size_t evaluation_number)
	Computes and returns whether the given evaluation number triggers the second condition of the logging based on the number of evaluations. More...

static int	coco_observer_evaluations_trigger (coco_observer_evaluations_t *evaluations, const size_t evaluation_number)
	Returns 1 if any of the two triggers based on the number of evaluations equal 1 and 0 otherwise. More...

static void	coco_observer_evaluations_free (coco_observer_evaluations_t *evaluations)
	Frees the given evaluations object. More...

Detailed Description

Definitions of functions regarding COCO observers.

Macro Definition Documentation

#define COCO_MAX_EVALS_TO_LOG 1000

The maximum number of evaluations to trigger logging.

Note: This is not the maximal evaluation number to be logged, but the maximal number of times logging is triggered by the number of evaluations.

Function Documentation

coco_observer_t* coco_observer	(	const char *	observer_name,
		const char *	observer_options
	)

Constructs a COCO observer.

Currently, three observers are supported:

"bbob" is the observer for single-objective (both noisy and noiseless) problems with known optima, which creates *.info, *.dat, *.tdat and *.rdat files and logs the distance to the optimum.
"bbob-biobj" is the observer for bi-objective problems, which creates *.info, *.dat and *.tdat files for the given indicators, as well as an archive folder with *.adat files containing nondominated solutions.
"toy" is a simple observer that logs when a target has been hit.

Parameters

observer_name A string containing the name of the observer. Currently supported observer names are "bbob", "bbob-biobj", "toy". Strings "no_observer", "" or NULL return NULL.

observer_options

A string of pairs "key: value" used to pass the options to the observer. Some observer options are general, while others are specific to some observers. Here we list only the general options, see observer_bbob, observer_biobj and observer_toy for options of the specific observers.

"result_folder: NAME" determines the folder within the "exdata" folder into which the results will be output. If the folder with the given name already exists, first NAME_001 will be tried, then NAME_002 and so on. The default value is "default".
"algorithm_name: NAME", where NAME is a short name of the algorithm that will be used in plots (no spaces are allowed). The default value is "ALG".
"algorithm_info: STRING" stores the description of the algorithm. If it contains spaces, it must be surrounded by double quotes. The default value is "" (no description).
"number_target_triggers: VALUE" defines the number of targets between each 10**i and 10**(i+1) (equally spaced in the logarithmic scale) that trigger logging. The default value is 100.
"target_precision: VALUE" defines the precision used for targets (there are no targets for abs(values) < target_precision). The default value is 1e-8.
"number_evaluation_triggers: VALUE" defines the number of evaluations to be logged between each 10**i and 10**(i+1). The default value is 20.
"base_evaluation_triggers: VALUES" defines the base evaluations used to produce an additional evaluation-based logging. The numbers of evaluations that trigger logging are every base_evaluation * dimension * (10**i). For example, if base_evaluation_triggers = "1,2,5", the logger will be triggered by evaluations dim*1, dim*2, dim*5, 10*dim*1, 10*dim*2, 10*dim*5, 100*dim*1, 100*dim*2, 100*dim*5, ... The default value is "1,2,5".
"precision_x: VALUE" defines the precision used when outputting variables and corresponds to the number of digits to be printed after the decimal point. The default value is 8.
"precision_f: VALUE" defines the precision used when outputting f values and corresponds to the number of digits to be printed after the decimal point. The default value is 15.
"precision_g: VALUE" defines the precision used when outputting constraints and corresponds to the number of digits to be printed after the decimal point. The default value is 3.
"log_discrete_as_int: VALUE" determines whether the values of integer variables (in mixed-integer problems) are logged as integers (1) or not (0 - in this case they are logged as doubles). The default value is 0.

Returns: The constructed observer object or NULL if observer_name equals NULL, "" or "no_observer".

static coco_observer_t* coco_observer_allocate	(	const char *	result_folder,
		const char *	observer_name,
		const char *	algorithm_name,
		const char *	algorithm_info,
		const size_t	number_target_triggers,
		const double	target_precision,
		const size_t	number_evaluation_triggers,
		const char *	base_evaluation_triggers,
		const int	precision_x,
		const int	precision_f,
		const int	precision_g,
		const int	log_discrete_as_int
	)

static

Allocates memory for a coco_observer_t instance.

static coco_observer_evaluations_t* coco_observer_evaluations	(	const char *	base_evaluations,
		const size_t	dimension
	)

static

Creates and returns a structure containing information on triggers based on evaluation numbers.

The numbers of evaluations that trigger logging are any of the two:

every 10**(exponent1/number_of_triggers) for exponent1 >= 0
every base_evaluation * dimension * (10**exponent2) for exponent2 >= 0

Note: The coco_observer_evaluations_t object instances need to be freed using the coco_observer_evaluations_free function!

Parameters

base_evaluations Evaluation numbers formatted as a string, which are used as the base to compute the second trigger. For example, if base_evaluations = "1,2,5", the logger will be triggered by evaluations dim*1, dim*2, dim*5, 10*dim*1, 10*dim*2, 10*dim*5, 100*dim*1, 100*dim*2, 100*dim*5, ...

static void coco_observer_evaluations_free ( coco_observer_evaluations_t * evaluations )

static

Frees the given evaluations object.

static int coco_observer_evaluations_trigger	(	coco_observer_evaluations_t *	evaluations,
		const size_t	evaluation_number
	)

static

Returns 1 if any of the two triggers based on the number of evaluations equal 1 and 0 otherwise.

The numbers of evaluations that trigger logging are any of the two:

every 10**(exponent1/number_of_triggers) for exponent1 >= 0
every base_evaluation * dimension * (10**exponent2) for exponent2 >= 0

static int coco_observer_evaluations_trigger_first	(	coco_observer_evaluations_t *	evaluations,
		const size_t	evaluation_number
	)

static

Computes and returns whether the given evaluation number triggers the first condition of the logging based on the number of evaluations.

The second condition is: evaluation_number == 10**(exponent1/number_of_triggers)

static int coco_observer_evaluations_trigger_second	(	coco_observer_evaluations_t *	evaluations,
		const size_t	evaluation_number
	)

static

Computes and returns whether the given evaluation number triggers the second condition of the logging based on the number of evaluations.

The second condition is: evaluation_number == base_evaluation[base_index] * dimension * (10**exponent2)

void coco_observer_free ( coco_observer_t * observer )

Frees the given observer.

const char* coco_observer_get_result_folder ( const coco_observer_t * observer )

Returns result folder name, where logger output is written.

Get the result folder name, which is a unique folder name constructed from the result_folder option.

Parameters

observer The COCO observer, whose logger may be wrapping a problem.

Returns: The result folder name, where the logger writes its output.

static coco_observer_targets_t* coco_observer_targets	(	const size_t	number_of_targets,
		const double	precision
	)

static

Creates and returns a structure containing information on targets.

Parameters

number_of_targets	The number of targets between 10(i/n) and 10((i+1)/n) for each i.
precision	Minimal precision of interest.

static int coco_observer_targets_trigger	(	coco_observer_targets_t *	targets,
		const double	given_value
	)

static

Computes and returns whether the given value should trigger logging.

coco_problem_t* coco_problem_add_observer	(	coco_problem_t *	problem,
		coco_observer_t *	observer
	)

Adds an observer to the given problem.

Wraps the observer's logger around the problem if the observer is not NULL and invokes the initialization of this logger.

Parameters

problem	The given COCO problem.
observer	The COCO observer, whose logger will wrap the problem.

Returns: The observed problem in the form of a new COCO problem instance or the same problem if the observer is NULL.

coco_problem_t* coco_problem_remove_observer	(	coco_problem_t *	problem,
		coco_observer_t *	observer
	)

Removes an observer from the given problem.

Frees the observer's logger and returns the inner problem.

Parameters

problem	The observed COCO problem.
observer	The COCO observer, whose logger was wrapping the problem.

Returns: The unobserved problem as a pointer to the inner problem or the same problem if the problem was not observed.

Data Structures

Macros

Functions

Detailed Description

Macro Definition Documentation

Function Documentation